Abstract:Network data in real-world tends to be error-prone due to incomplete sampling, imperfect measurements, etc.; this in turn results in inaccurate results when performing network analysis or modeling, such as node classification and link prediction, on these flawed networks. In this paper, we aim to reconstruct a reliable network from a flawed, undirected, unweighted network, a process referred to network enhancement. More specifically, network enhancement aims to detect the noisy links that are observed in the network but should not exist in the real world, as well as to predict the missing links that do indeed exist in the real world yet remain unobserved. While some attempts have been made to detect either noisy links or missing links, few of these works have considered unifying these two tasks, even though they are inter-dependent and capable of mutually boosting each others’ performance. In this paper, we therefore propose E-Net, an end-to-end graph neural network model, to leverage the mutual influence of these two tasks in order to achieve both goals more effectively. On one hand, detecting noisy links can benefit the performance of missing link prediction, while on the other hand, predicting missing links can provide indirect supervision for detecting noisy link detection when the labels of these noisy links are unavailable. Moreover, by proposing a subgraph extraction mechanism based on random walk with restart, the model can be scaled up to large networks and is able to preserve the local and global structural characteristics. The experimental results on several types of large networks demonstrate that the proposed model obtains an improvement of 10.7 percent on average in terms of F1 for predicting missing links, along with an average of 3.7 percent improvement in terms of precision for detecting noisy links compared with the state-of-the-art baselines.

Improved Graph Autoencoder for Network Reconstruction

Robust Network Enhancement from Flawed Networks.

Graph autoencoder for directed weighted network

Network Embedding Algorithm Taking in Variational Graph AutoEncoder

AENEA: A novel autoencoder-based network embedding algorithm

Exploring the Representational Power of Graph Autoencoder

Enhancing the Network Embedding Quality with Structural Similarity

Semi-AttentionAE: An Integrated Model for Graph Representation Learning

Online Dynamic Network Embedding

Node Level Graph Autoencoder: Unified Pretraining for Textual Graph Learning

Graph Autoencoder with Preserving Node Attribute Similarity

Masked Graph Autoencoder with Non-discrete Bandwidths

Stabilizing and Enhancing Link Prediction through Deepened Graph Auto-Encoders

Feature-Dependent Graph Convolutional Autoencoders with Adversarial Training Methods

Recovering Missing Node Features with Local Structure-based Embeddings

Advances in Knowledge Discovery and Data Mining - 23rd Pacific-Asia Conference, PAKDD 2019, Macau, China, April 14-17, 2019, Proceedings, Part I.

Deep Recursive Network Embedding With Regular Equivalence

Auditing Network Embedding: an Edge Influence Based Approach

Network embedding based on high-degree penalty and adaptive negative sampling

Multi-View Graph Autoencoder for Unsupervised Graph Representation Learning

A general deep-learning approach to node importance identification